<<

This document is downloaded from DR‑NTU (https://dr.ntu.edu.sg) Nanyang Technological University, Singapore.

Broadband strip‑line ferromagnetic resonance of soft magnetic CoFeTaZr patterned thin films

Jin, Tian Li; Nongjai, R.; Asokan, K.; Ghosh, A.; Aparnadevi, M.; Suri, P.; Piramanayagam, S. N.; Gupta, Surbhi; Kumar, Durgesh

2018

Gupta, S., Kumar, D., Jin, T. L., Nongjai, R., Asokan, K., Ghosh, A., et al. (2018). Broadband strip‑line ferromagnetic resonance spectroscopy of soft magnetic CoFeTaZr patterned thin films. AIP Advances, 8(5), 056125‑. https://hdl.handle.net/10356/86295 https://doi.org/10.1063/1.5007943

© 2018 The Author(s) (published by American Institute of Physics). All article content, except where otherwise noted, is licensed under a Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

Downloaded on 05 Oct 2021 15:01:54 SGT AIP ADVANCES 8, 056125 (2018)

Broadband strip-line ferromagnetic resonance spectroscopy of soft magnetic CoFeTaZr patterned thin films S. Gupta,1 D. Kumar,1 T. L. Jin,1 R. Nongjai,2 K. Asokan,2 A. Ghosh,3 M. Aparnadevi,4 P. Suri,4 and S. N. Piramanayagam1,a 1School of Physical and Mathematical Sciences, Nanyang Technological University, 21 Nanyang link, 637371, Singapore 2Materials Science Division, Inter-University Accelerator Centre, New Delhi 11006, India 3Data Storage Institute, A∗STAR (Agency for Science, Technology and Research), 2 Fusionopolis Way, 08-01 Innovis, 138634, Singapore 4Heraeus Materials, 1 Tuas South Street 3, Singapore 569059 (Presented 7 November 2017; received 4 October 2017; accepted 8 November 2017; published online 18 January 2018)

In this paper, magnetic and dynamic properties of compositionally pat- terned Co46Fe40Ta9Zr5 thin films are investigated. A combination of self-assembly and ion-implantation was employed to locally alter the composition of Co46Fe40Ta9Zr5 thin film in a periodic manner. 20 keV O+ and 60 keV N+ ions were implanted at different doses in order to modify the magnetization dynamic properties of the samples in a controlled fashion. Magnetic hysteresis loop measurements revealed sig- nificant changes in the coercivity for higher influences of 5 × 1016 ions per cm2. In particular, N+ implantation was observed to induce two phase formation with high and low coercivities. Broadband strip-line ferromagnetic resonance spectroscopy over wide range of frequency (8 – 20 GHz) was used to study the magnetization dynamics as a function of ion-beam dosage. With higher fluences, damping constant showed a continuous increase from 0.0103 to 0.0430. Such control of magnetic prop- erties at nano-scale using this method is believed to be useful for spintronics and device applications. © 2018 Author(s). All article content, except where otherwise noted, is licensed under a Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/). https://doi.org/10.1063/1.5007943

I. INTRODUCTION Investigation of soft amorphous magnetic films is a topic of interest for magnetic recording, spin electronics, thin film transformers, magnetic inductors and microwave communication. In the field of magnetic recording, the soft magnetic films are useful in the writing heads and in the perpendicular magnetic media for the purpose of writing operation.1,2 In the field of microwave communication, they are ideal candidates for high frequency harmonic noise suppression due to the ability of their magnetization to precess at a high frequency (GHz range).3,4 In particular, CoZr based system has been investigated extensively for microwave communication in the past.3–6 In such a system, the addition of elements such as Nb, Ta etc. is carried out to reduce the magnetostriction to nearly zero and to stabilize the amorphous phase of the CoZr-based magnetic material.1 In particular Y. Endo et al.4 and G. Hao et al.5 studied the effect of composition variation and electrical pulse annealing respectively, on high frequency magnetic properties. Interestingly investigations on CoZr based sys- tems, thus far, have mainly been focused on continuous films. Since the magnetization permeability largely depends on the ferromagnetic resonance (FMR) frequency, lamination as well as compos- ites of CoZr and oxides has also been investigated as alternative to increase the FMR frequency.7–9

aCorresponding author: [email protected]

2158-3226/2018/8(5)/056125/6 8, 056125-1 © Author(s) 2018

056125-2 Gupta et al. AIP Advances 8, 056125 (2018)

In the case of laminations, the magnetic and non-magnetic layers are stacked one over the other to improve the high-frequency performance of the stack. Recently T. N. Kołtunowicz et al.9 discussed the FMR of “FeCoZr core – Al2O3 oxide shell” nanoparticles and observed the composition dependent behavior. As an alternative to lamination, we have attempted to investigate the high frequency performance by local patterning in the lateral direction. In contrast to a continuous Co-based amorphous film with laminations, the composition that is modulated periodically in the lateral direction is expected to perform differently. Simulation studies on patterned soft magnetic underlayer has been proposed to improve recording performance of magnetic media.10 However, there are not many experimental results based on nanometer scale patterning.11 For example, Ikeda et al.11 explored high frequency properties by designing micro-patterned CoNbZr films of different feature size for band stop filter application. In order to understand the effect of patterning on a nano-scale, by inducing periodic com- position modulation, we employed a combination of self-assembly and high-energy ion-implantation. Self-assembly based on di-block copolymers can be effectively used to achieve periodic patterns of varying pitch over a magnetic layer.12 Ion-implantation of non-magnetic species (such as O and N) was carried out through the self-assembled mask patterns, in order to modulate the composition of present Co-based amorphous films periodically. Such patterned Co-based amorphous films offer an interesting system to study for high frequency applications. In addition, we have chosen a CoFeTaZr system instead of commonly studied CoZr, with a purpose of achieving higher magnetization (due to the presence of CoFe), relative softness (i.e., low coercitive field Hc) and their low magnetization precessional damping. The understanding based on hysteresis loop and FMR investigations on such compositionally patterned CoFeTaZr films are reported in this paper.

II. EXPERIMENTAL DETAILS Soft CoFeTaZr magnetic thin films were deposited on silicon substrates by DC magnetron sput- tering using a 2-inch Co46Fe40Ta9Zr5 alloy target. A high pressure deposition (50 mTorr) was used to achieve a granular CoFeTaZr film, instead of a continuous film.12 Surface morphology of films were scanned by Atomic Force Microscopy. Figure 1(a) shows the microstructure of as-deposited pristine films in scanned area of 1 µm × 1 µm. The AFM shows the granular nature of CoFeTaZr film as intended, where cluster size is found to be ranging from 10 to 60 nm. Self-assembly technique of di-block copolymer i.e. polystyrene (PS)-polydimethylsiloxane (PDMS) was employed to fabricate the patterned mask of oxidized PDMS/SiO2 spherical dots on CoFeTaZr magnetic thin films as shown in Figure 1(b). For this purpose, commercially available PS-PDMS (88:22 weight fractions) di-block copolymer of molecular weight 54 kg/mol was mixed in Toluene and spin coated over magnetic thin films. Then solvent annealing was performed to enable the complete micro-phase separation between PS and PDMS.12 Subsequently the di-block copolymer film was reactive ion etched to remove the PS matrix as well as oxidize the PDMS nanostructure. AFM micrograph in Fig. 1(b) shows the hexagonally coordinated sphere morphology (average feature size ∼ 35 nm ± 2 nm) in PDMS mask. For implantation, 20 keV of O+ and 60 keV of N+ were chosen to incorporate the maximum fluence

FIG. 1. AFM micrograph of (a) as-deposited CoFeTaZr thin film and (b) Pattern mask of oxidized PDMS/SiO2 spherical dots on CoFeTaZr magnetic thin film and (c) Ion profile obtained from TRIM on implanting 20 keV O+ and 60 keV N+ in CoFeTaZr thin film. 056125-3 Gupta et al. AIP Advances 8, 056125 (2018) at about half the film thickness and to minimize the ion-beam mixing at the CoFeTaZr/Si interface. The range of implanted ions in the different layers was verified using the Transport of Ions in Mat- ter (TRIM) program from the SRIM software.13 The implantation profile obtained from TRIM for 20 keV of O+ and 60 keV of N+ are shown in figures 1(c). It should be noted that maximum implan- tation fluence is centered either before or half the film thickness. However, slight spread of ions can be seen in substrate in addition to the targeted layer in case of N+ implantation. The implantation was performed at two fluences, i.e. 1 × 1016 and 5 × 1016 ions per cm2 through oxidized mask of PDMS/SiO2 spherical dots. In-plane magnetic hysteresis loops were traced using vibrating sample magnetometer. Ferromag- netic resonance (FMR) in field sweep mode was recorded using a vector network analyzer, connected to co-planar waveguide (CPW) using 50 ohm coaxial cables and high-frequency connectors. The samples were mounted in flip-chip geometry such that external static magnetic field is applied in the in-plane direction.

III. RESULTS AND DISCUSSION Figures 2(a) and 2(b) show the hysteresis loops measured along the in-plane direction before and after the 20 keV O+ and 60 keV N+ implantation respectively. As shown in the inset, pris- tine CoFeTaZr sample showed isotropic soft magnetic behavior with high saturation magnetization Ms = 1021 emu/cc and a coercive field (Hc) of 15 Oe. This value of Hc is smaller in comparison to that in crystalline CoFe thin films, and is good for high frequency application.3,15 It may be seen from figure 2(a) that a low dosage of 1× 1016 O+ ions/cm2 did not induce significant changes in magnetic behavior as a marginal increase in remnant magnetization (Mr) is observed. However, dosage of 16 + 2 5× 10 O ions/cm led to visible increment in Mr as well as Hc attributed to compositionally modified regions, acting as pinning centers and thereby increasing the coercivity.14 Surprisingly, 60 keV N+ implantation led to contrasting observations (shown in fig. 2(b)). In case 16 + 2 of small fluence of 1× 10 N ions/cm ,Hc first decreased to 9 Oe while Mr showed a significant 16 + 2 upsurge. At a higher fluence of 5× 10 N ions/cm , an evident enhancement in Hc with the highest value of 80 Oe is observed with a drop in Mr. Similar trend has also been observed for Ne, Xe implanted 13 FeCo thin films in literature. Also, the squareness factor (Mr/Ms) is found to be improved with high implantation dosage irrespective of the ion species. Magnetization dynamics in the sample was induced by a microwave Oersted (dynamic) magnetic field hrf, which is generated by a microwave current (0 dBm, 8 - 20 GHz) flowing through the signal line of the co-planar waveguide and is maintained to be perpendicular to external magnetic field, as shown in inset of Figure3. Figure 3(a) shows the typical FMR spectra for pristine CoFeTaZr thin film, recorded for different applied frequencies (8 GHz to 20 GHz). The external magnetic field was swept up to 4000 Oe. Here scattering parameter S21 corresponds to ratio of absorbed microwave power at port 2 to the incident fixed microwave power of 0 dBm at port 1. The characteristic spectrum showed a complex line shape signal for all the applied frequency ranges where continuous decrease in the

FIG. 2. M-H Hysteresis loop CoFeTaZr thin films implanted with different fluences of (a) 20 keV O+ and (b) 60 keV N+ respectively. Inset shows the isotropic behavior of pristine sample. 056125-4 Gupta et al. AIP Advances 8, 056125 (2018)

FIG. 3. (a) Field sweep FMR spectra of pristine CoFeTaZr thin film, (b) Resonance field (open circle) as obtained from experiments and a curve fit by Kittel formula (solid line) and (c) Linear fit of field linewidth (∆H) versus applied frequency. Inset shows the sample location around the CPW in flip chip geometry. absorption peak intensity is attributed to high frequency power losses. The obtained FMR signals as a function of applied magnetic field were fitted with sum of derivative of symmetric and antisymmetric 16 Lorentzian functions as given in equation1 to estimate the linewidth ( ∆H) and resonance field (Hres). 2 ! X (∆H) ∆H(H − Hres) S12 (H) = L + D (1) n 2 2 2 2 (∆H) + (H − Hres) (∆H) + (H − Hres) where L and D is the respective amplitude of symmetric and antisymmetric Lorentzian functions. The submission over the symmetric and antisymmetric component is performed with “n” defin- ing the number of ferromagnetic elements in the system. The different values of resonance field (Hres) and field line-widths (∆H) thus obtained are plotted in Figure 3(b) and 3(c) respectively. 17 In Figure√ 3(b), the plot of applied frequency (ω) versus Hres is further fitted by Kittel formula, ω = γµO Hres(Hres + 4πMeff) to determine the effective magnetization (4πMeff) to be 1.25 T. Here γ = gµB/h is the , g is the Lande’s factor, µB is the Bohr magnetron and h is the Planck constant. The other fundamental quantity that characterizes high-frequency response of a ferromagnetic material is the damping, which basically tells about intrinsic relaxation mechanism of the precessional motion of magnetization vector/dynamics. The dimensionless Gilbert damping coefficient (α) in addition to extrinsic magnetic inhomogeneity (∆Ho) was determined from ∆H, according to the 15 equation µO (∆H- ∆Ho) = αω/γ. Here linear increase in ∆H with applied frequency as seen in figure 3(c) indicates the Gilbert-type damping in CoFeTaZr thin films with damping coefficient of 0.0103 which is found to be in agreement with the damping constant of CoFeZr.3 In addition, zero- frequency contribution, i.e. the intercept ∆Ho corresponding to the inhomogeneous broadness is found to be 2.89 mT. Similar measurements were also performed for 20 keV O+ implanted samples. No significant change could be observed for low dosage of 1× 1016 ions/cm2. Figure 4(a) shows the FMR spectrum of sample with 5× 1016 O+ ions/cm2 implantation for selected frequencies.

FIG. 4. Typical FMR spectra of CoFeTaZr thin film implanted with higher 5 × 1016 ions/cm2 fluence of (a) 20 keV O+ and (b) 60 keV N+ ion. Inset shows the M-H loop of N+ implanted film. 056125-5 Gupta et al. AIP Advances 8, 056125 (2018)

TABLE I. The magnetic and high frequency parameter of pristine and ion-implanted samples.

Species and Dose Hc (Oe) Lande g factor 4πMeff (T) α ∆Ho (mT) Unimplanted 15 2.18 1.25 0.0103 2.89 1 × 1016 O+ ions/cm2 13 2.18 1.30 0.0112 1.17 5 × 1016 O+ ions/cm2 40 2.18 1.29 0.0137 2.41 1 × 1016 N+ ions/cm2 9 2.18 1.18 0.0184 0.73 5 × 1016 N+ ions/cm2 80 2.18 1.35 0.0430 1.10

Broadness in all microwave absorption spectra is quite evident in comparison to pristine sample. In contrast, two resonance peaks were observed in FMR spectrum of 5 × 1016 N+ ions/cm2 implanted CoFeTaZr samples. We attribute this to the possibility of two magnetic phases in the films after implan- tation (Figure 4(b)). This is further supported by the observation of two reversals in the M-H loops of 5 × 1016 N+ ions/cm2 implanted CoFeTaZr thin film samples (shown in inset). In case of N+ implanted sample, the fitting of complex FMR signal has been performed using equation1 and keeping n = 2 owing to double peak spectra. Typical magnetic and high frequency properties of pristine as well as implanted samples extracted from the M-H hysteresis and FMR measurements are tabulated in TableI. There is clear increase in Hc for higher fluences of implantation while damping showed the steadily increase with implantation fluence. It may be noted that N+ implantation has induced larger impact on the magnetic properties in comparison to O+ implantation. Owing to higher energy of the 60 keV N+ implanted ions, we proposed the significant change is may be due to enhanced scattering and modification in the spin- orbit interaction. As far as Gilbert damping is concerned, we observed monotonic increase in the damping with increase in ion-dose, found to be matching with literature.18 It is important to mention that higher damping is also desirable for increasing the switching rate to ensure coherent reversal in magnetic elements in magnetic storage devices.18 However, the estimated homogeneous linewidth contribution is observed to be scattered and requires further understanding. The results from this study indicate that the proposed method of localized patterning is an alternative way to tailor the magnetization dynamics of soft magnetic films.

IV. CONCLUSION In conclusion, static as well as dynamic magnetic properties of patterned CoFeTaZr films deposited by DC magnetron sputtering on silicon substrates were investigated using VSM and FMR. We have shown, by means of 20 keV O+ and 60 keV N+ implantation, that soft magnetic properties of CoFeTaZr thin film can be tuned over a large range. Implanted O+ ions increased the coercivity as well as the remnant magnetization. Samples with a higher fluence of 5 × 1016 N+ ions/cm2 implantation showed two magnetic phases, resulting in a significant enhancement in the coercivity and damping constant as well. This combined methodology of self-assembly and ion-implantation for local control of damping and static magnetic properties is expected to be useful in the field of spintronics as well as microwave communication.19

ACKNOWLEDGMENTS The authors gratefully acknowledge MOE AcRF Tier1 grant RG163/15 and NRF-IIP grant ((NRF2015-IIP003-001) for the partial financial support. We duly acknowledge the low energy ion beam facility (LEIBF) at Inter University Accelerator Centre (IUAC), New Delhi. 1 J. Shi, Y. Yang, H. K. Tan, S. N. Piramanayagam, C. B. Lim, H. L. Seet, S. L. Ho, and J. F. Hu, Phys. Stat. Sol. PSS-RRL 11(2), 1600341 (2017). 2 S. N. Piramanayagam and K. Srinivasan, J. Magn. Magn. Mater. 321(6), 485 (2009). 3 C. L. Graet,¨ D. Spenato, N. Beaulieu, D. T. Dekadjevi, J. P. Jay, S. P. Pogossian, B. W. Fonrose, and J. B. Youssef, Euro. Phys. Lett. 115, 17002 (2016). 4 Y. Endo, T. Ito, T. Miyazaki, Y. Shimada, and M. Yamaguchi, J. Appl. Phys. 117, 17A330 (2015). 5 G. Hao, H. Zhang, and X. Tang, J. Appl. Phys. 113, 17A341 (2013). 6 D. W. Lee and S. X. Wang, J. Appl. Phys. 99, 08F109 (2006). 056125-6 Gupta et al. AIP Advances 8, 056125 (2018)

7 A. E. Ghazaly, R. M. White, and S. X. Wang, J. Appl. Phys. 117, 17E502 (2015). 8 G. Hao, D. Zhang, X. Tang, and H. Zhang, J. Appl. Phys. 115, 17A337 (2014). 9 T. N. Kołtunowicz, P. Zukowski, J. Sidorenko, V. Bayev, J. A. Fedotova, M. Opielak, and A. Marczuk, J. Magn. Magn. Mater. 421, 98 (2017). 10 Y. Hijazi, R. Ikkawi, N. Amos, A. Lavrenov, D. Doria, N. Joshi, R. Chomko, D. Litvinov, and S. Khizroev, IEEE Trans. Magn. 42(10), 2375 (2006). 11 S. Ikeda, T. Nagae, Y. Shimada, K. H. Kim, and M. Yamaguchi, J. Appl. Phys. 99, 08P507 (2006). 12 R. A. Griffiths, A. Williams, C. Oakland, J. Roberts, A. Vijayaraghavan, and T. Thomson, J. Phys. D: Appl. Phys. 46, 503001 (2013). 13 D. Kumar, S. Gupta, T. Jin, R. Nongjai, K. Asokan, and S. N. Piramanayagam, IEEE Magn. Lett. 14 S. N. Piramanayagam, J. Appl. Phys. 102, 011301 (2007). 15 R. Gupta, K. H. Han, K. P. Lieb, G. A. Muller,¨ P. Schaaf, and K. Zhang, J. Appl. Phys. 97, 073911 (2005). 16 M. Weiler, J. M. Shaw, H. T. Nembach, and T. J. Silva, IEEE Mag. Lett. 5, 3700104 (2014). 17 S. Gupta, R. Medwal, D. Kodama, K. Kondou, Y. Otani, and Y. Fukuma, Appl. Phys. Lett. 110, 022404 (2017). 18 J. A. King, A. Ganguly, D. M. Burn, S. Pal, E. A. Sallabank, T. P. A. Hase, A. T. Hindmarch, A. Barman, and D. Atkinson, Appl. Phys. Lett. 104, 242410 (2014). 19 R. Sbiaa and S. N. Piramanayagam, Phys. Stat. Sol. PSS-RRL 1700163 (2017).